Our study has clearly demonstrated that zebrafish can be a suitable vertebrate model for studying human craniofacial development and disease, and the finding of an important role of tgf
Trang 1GENE IN ZEBRAFISH
CHEAH SIEW HONG FELICIA (BSc Biochemistry/Microbiology (Hons.), University of
Trang 2Acknowledgements ……… ……. iii
List of Figures ……… iv
List of Tables …….……… vi
List of Abbreviations ……… vii
List of Publications ……….…… x
Chapter 1 Introduction 1
1.1 Background/Significance ……… ……… 1
1.2 Literature Reviews ……… ……… 3
1.2.1 Transforming Growth Factor b3 ……… … 3
Overview of transforming growth factor b family ……… 3
TGF b ligands ……… 3
TGF b receptors ……… ……… 5
Smad proteins ……… ……. 7
Recognition of Smad by the activated receptor complex ……… … 9
Mechanism of Smad phosphorylation and activation ……… 10
Mechanisms of TGF b signalling from cell membrane to the nucleus ……… 10
TGF b Subfamily and its isoforms ……… 11
Latent TGF b , latent TGF b binding protein, and bone … 11
Perturbation of TGF b signalling in human diseases ………. 13
TGF b 3, and expression patterns during early embryogenesis and in adult tissues 13
TGF b 3 knockout mice studies ……… 17
TGF b 3 mutations and diseases ……… 19
1.2.2 Zebrafish: An Animal Model for Craniofacial Development and Disease 21
Overview of the zebrafish system ……… 21
Advantages and disadvantages of zebrafish ……… 21
Disease modelling in zebrafish ……… 22
Zebrafish as the animal model for craniofacial development and disease ………… 25
1.2.3 Development of Pharyngeal Arches ……… 27
Overview of pharyngeal arches ………. 27
Patterning the pharyngeal arches ……… 27
Trang 3Functions of the notochord in vertebrate development ……….……… 37
Relationship between notochord and cartilage ……….………. 39
1.2.5 Cardiac Development in Zebrafish ……… 41
Overview of zebrafish heart ……… 41
Formation of the heart ……… … 41
1.2.6 Specific Aims ……… 47
Chapter 2 Materials and Methods ……… 48
2.1 Zebrafish maintenance ……… 48
2.2 Verification of RZPD clone ………. 48
2.3 DNA sequencing and cDNA analysis ……… 49
2.4 Genomic characterization ……….……… 49
2.5 Promoter analysis ……….… 50
2.6 Embryonic expression pattern by RTPCR analysis ………… ………. 53
2.7 Embryonic expression pattern by in situ RNA hybridization analysis …… … 53
2.8 Cryosection analysis ……….… 55
2.9 Microinjection ……… ……… ……….……… 56
2.10 Knockdown studies using morpholino antisense oligonucleotides …… ………. 56
2.11 RTPCR analysis for detection of morphant transcripts ……… ……….… …. 57
2.12 Realtime PCR analysis for efficacy of i1e2 splice modifying MO ……… 57
2.13 Overexpression analysis ……….………… 58
2.14 Alcian blue cartilage staining ……….……….………. 60
2.15 Alizarin red bone staining ……… ….…… 61
Trang 42.17 Imaging and processing ……… 62
Chapter 3 Results ……… 63
3.1 Identification of Zebrafish tgf b3 cDNA ………. 63
3.2 Genomic Organization of Zebrafish tgf b3 Gene ……….… 69
3.3 Tgfb3 Promoter Analysis ………. 71
3.4 Genetic Mapping of Zebrafish tgf b3 ……… 75
3.5 Embryonic Developmental Expression Pattern ……….…… 77
3.6 Inhibition of Zebrafish tgf b3 with Splice Modifying Morpholino Ô ……… ……84
3.7 Efficacy Assessment of Knockdown Morphant Phenotypes ……… 89
3.8 Craniofacial Cartilage Phenotypes in tgf b3 Knockdown Morphant Embryos …92 3.9 Knockdown of tgf b3 Gene and Bone Development ……… 97
3.10 Knockdown of tgf b3 Gene and Notochord Development ……….…… 99
3.11 Undulating Notochord and Heart Field Domain in tgf b3 Morphants …… … 102
3.12 Knockdown of tgf b3 Gene and Heart Development ……….…….… 104
3.13 Overexpression of tgf b3 in Zebrafish ……….… 110
Capped sense mRNA synthesis and validation of mRNA integrity ……… 110
Overexpression of tgf b 3 and cartilage development ……… 112
Overexpression of tgf b 3 and notochord development ………. 112
Overexpression of tgf b 3 and cardiac development ………. 112
Chapter 4 Discussion ……….…. 118
4.1 Genomic Analyses of tgf b3 ……… ………. 118
4.2 Functional Analyses of tgf b3 ………. 119
Embryonic expression of tgf b 3 ……… 119
Trang 5Relationship between tgf b 3 and cardiac development ………… ………. 128
Dosage sensitive and tissue specific effects of tgf b 3 ……… 132
Chapter 5 Conclusions 133
Chapter 6 References ……… … 135
Appendix ………. 148
(A) Medium Preparation
(B) Published articles
Trang 6biological processes and is involved in mammalian pulmonary and craniofacial
development. Homologs of human TGF b 3 have been identified in several vertebrate
species. A cDNA clone of zebrafish tgf b 3, consisting of a 271 bp 5’ untranslated region,
a 1233 bp open reading frame that encodes a predicted 410 amino acid peptide, and a 527
bp 3’ untranslated region was sequenced. Using 5’ rapid amplification of cDNA ends, the transcription start site of this gene was determined to lie an additional 29 nucleotides upstream. This gene is composed of seven exons and maps to a segment of linkage group
protein (Sp1) and two TATA binding protein (TBP) transcription factor binding sites were identified in the putative promoter segment upstream of the transcription start site.
Comparative alignment analysis revealed a high degree of tgf b 3 nucleotide and amino
acid identity between zebrafish and other species, including a complete conservation of the cysteine knot structure that facilitates proteinprotein interaction. Also, 9 out of 10 amino acid residues critical for ligand/receptor binding in human TGFb3 are conserved in zebrafish, suggesting a high degree of functional conservation even in lower vertebrates.
Zebrafish tgf b 3 expression was first detected in the notochord (10 somite to highpec
stage), subsequently in the developing pharyngeal arch and neurocranial cartilage (18 somite to protruding mouth stage), lens and heart (21 somite to protruding mouth stage), and pectoral fins (prim25).
Trang 7Both cartilage staining and molecular marker analysis results showed that morphant larvae had reduced pharyngeal arches, neurocranial cartilage and pectoral fin,
confirming that tgf b 3 is involved in the formation of cartilages of the pharyngeal arches,
neurocranium, and pectoral fin appendages. The quadrate bone that forms the main part
of the upper jaw skeleton and lateral part of the larval palate was also absent in the
morphants. This observation is reminiscent of the cleft palate phenotype reported in the
Tgf b 3/ null mice, suggesting that the role of TGF b 3 in palatogenesis has been
conserved throughout vertebrate evolution. The opercle and the branchiostegal rays which form the key supportive components of the gill chamber of zebrafish were also
reduced in the morphants, suggesting that tgf b 3 is required for the proper assembly of the
gill chamber. Tgf b 3 also appears to be essential for the proper formation of the heart.
Our studies have revealed that loss of tgf b 3 expression affects the heart field formation,
cardiac cone formation, heart tube elongation, and heart tube looping in cardiac
morphogenesis. Tgf b 3 may regulate the cardiomyocytes population by limiting the
expansion of heart field domain in the midline via its effect on the notochord, and regulating the population of neural crest that would differentiate into cardiomyocytes.
Our study has clearly demonstrated that zebrafish can be a suitable vertebrate model for studying human craniofacial development and disease, and the finding of an
important role of tgf b 3 in cardiac development is novel and has not been reported in other
model organisms.
Trang 8I thank my supervisor, A/P Samuel Chong for his invaluable guidance and patience, Yi Zhou (The Children’s Hospital Zebrafish Genome Project Initiative, Boston, MA) for performing the RH mapping, Vladimir Korzh (Institute of Molecular and Cell Biology, Singapore) for invaluable advice and technical support, Bill Trevarrow (University of Oregon Zebrafish Facility, Eugene, OR) for providing breeding stocks of the AB line,
Karuna Sampath (Temasek Life Science Laboratories, Singapore) for nkx2.5 and ntl molecular markers, YiLin Yan (University of Oregon, Eugene, OR) for sox9a molecular marker, Monte Westerfield (University of Oregon, Eugene, OR) for dlx2 molecular
marker, Bonnie Ullmann (University of Oregon, Eugene, OR) for alizarin red staining protocol, Jin Ben and Gare Hoon Yeo for technical assistance and motivation, and lab members of A/P Chong’s lab for their constant encouragement and motivation. Last but not least, my beloved husband, Dennis Goh for his constant support, encouragement and
love during the arduous six years of my parttime graduate studies.
“The Sovereign LORD is my strength; he makes my feet like the feet of a deer, he enables
me to go on the heights.” Habakkuk 3:19
Trang 9List of Figures Fig. 1 The relationship between hindbrain, neural crest migration, and
Fig. 11 Zebrafish tgf b 3 expression in the notochord (AD) and lens (EH)
Fig. 12 Zebrafish tgf b 3 expression in the presumptive pharyngeal arch primordia,
pharyngeal arches, and neurocranial cartilage
Fig. 13 Zebrafish tgf b 3 expression in the pectoral fins (AC, F) and heart (DH)
Fig. 14 Inhibition of zebrafish tgf b 3 with splice modifying Morpholino Ô
Fig. 15 Relative quantitation of morphant transcripts in uninjected control and
Trang 10Fig. 21 Expression of cartilage marker sox9a in the pharyngeal arches,
neurocranial cartilage, and pectoral fins of 48hpf wildtype (WT) and
tgfβ3 morphant (tgfβ3 MO ) larvae
Fig. 22 Expression of neural crest marker dlx2 in 14somite stage wildtype (WT)
and tgfβ3 morphant (tgfβ3 MO ) embryos
Fig. 23 Pharyngeal bones development in knockdown tgfβ3 morphant larva
Fig. 24 Notochord phenotype in knockdown tgfβ3 morphant embryos
Fig. 25 Notochord crosssectional area in morphant embryos
Fig. 26 Loss of tgf b 3 affects the organization and posterior extension of the heart
field domain
Fig. 27 Cardiac cones formation in tgfb3 MO embryos
Fig. 28 Heart tubes elongation and looping in tgfb3 MO embryos
Fig. 29 Morphogenesis of Kupffer’s vesicle (KV) is affected in tgfb3 MO embryos
Fig. 30 Synthesis and validation of tgf b 3 capped mRNA used for overexpression
studies
Fig. 31 Cartilaginous head skeleton of a fourday old wildtype zebrafish larva
(WT) and a fourday old overexpressed larva (tgfβ3 OE )
Fig. 32 Notochord phenotype in overexpressed tgfβ3 (tgfβ3 OE ) embryos
Fig. 33 Cardiac cones formation in tgfb3 OE embryos
Fig. 34 Heart tubes elongation in tgfb3 OE embryos
Trang 151. Chong, S. S., Cheah, F. S. H., and Jabs, E. W. (2002) “Genes implicated in lip and
palate development” Chapter 3 of Cleft Lip and Palate: From Origin to Treatment
(Wyszynski, D.F.) Oxford University Press, USA.
2. Cheah, F. S. H., Jabs, E. W., and Chong, S. S. (2002). Expression and functional
315 (838).
3. Cheah, F. S. H., Jabs, E. W., and Chong, S. S. (2005). Genomic, cDNA, and
embryonic expression analysis of zebrafish transforming growth factor beta 3 (tgfb3).
Developmental Dynamics 232: 10211030.
4. Cheah, F. S. H., Jabs, E. W., and Chong, S. S. An essential role of tgf b 3 in
zebrafish pharyngeal arch and heart development (manuscript in preparation).
Trang 16Chapter 1 Introduction
1.1 Background/Significance
The anatomy of the head forms the most sophisticated part of the vertebrate body. Generally, the skull consists of two major entities, the neurocranium and the viscerocranium (Wilkie and MorrissKay, 2001). The neurocranium surrounds and protects the brain and sensory organs, whereas the viscerocranium forms the key skeleton of the face. Craniofacial malformation occurs when there is an abnormal regulation of normally coordinated tissue patterning, cell fate determination, and differentiation of specific cell types during early embryogenesis (Nuckolls et al., 1999). Craniofacial anomalies account for approximately a third of all congenital defects (Waheed, RM, 1998). These anomalies include disorders of the skull vault, oral clefts, and abnormalities associated with pharyngeal apparatus derivatives. The mechanisms that control the epithelialmesenchymal interactions which mediate facial outgrowth and morphogenesis are unclear. An understanding of the regulatory mechanisms involved will provide us with the fundamental knowledge about the genetic control of normal human craniofacial development, and more importantly it may aid us in the understanding on how a genetic mutation during embryogenesis can produce a particular
counseling and risk assessment, and is the necessary initial step for translational research towards improving patient management, and developing treatment.
As we know, craniofacial development is a highly controlled and complex process which begins with the formation of neural crest cells in the brain and their subsequent
Trang 17et al., 1998). The outgrowth of these facial primordia from buds of undifferentiated mesenchyme into intricate series of bones and cartilage structures, muscles and other tissues that form the face, involves a hierarchy of growth factors and their downstream transcription factors that regulate the expression of genes responsible for determining cell phenotypes during embryogenesis. In recent years, key important components of this hierarchy that regulate craniofacial morphogenesis have been identified. This includes
Trang 18in most of the human gastrointestinal cancers with microsatellites instability, the TGFb type II receptor is inactivated by mutation, and in nearly half of the pancreatic
Trang 19interactions between the monomers. In most cases, this dimer structure is further strengthened by the presence of an intersubunit disulfide bond. This ligand dimer forms
a complex with two type I and two type II receptors. A family of proteins known as ligand traps has been shown to modulate the accessibility of the ligands to the receptors
by blocking the ligand surfaces that are required to interact with type I and type II receptors (Groppe et al., 2002).
Generally, there are two distinct modes of ligandreceptor interaction (Shi and Massague, 2003). One of the modes is exemplified by members of the BMP subfamily while the other is represented by TGFβs and Activins subfamily. In the former mode, BMP ligands like BMP2 and BMP4 display a high affinity for the extracellular domains
on the BMPRI (type I BMP receptors) and a low affinity for the BMPRII (type II BMP receptors). These ligands bind directly to the ectodomain of the type I receptor first. This binding leads to subsequent interaction of ligandtype I receptor complex with type
II receptor. Crystal structure of BMP2 BMPRIA ectodomain complex shows that the receptor makes extensive contacts to BMP2 dimers by binding to the “wrist” epitope of the dimers (Kirsch et al., 2000). Most of the interactions are hydrophobic. The aromatic side chain of Phe85 of BMPRIA interacts with the hydrophobic pocket formed at the interface of the two BMP2 monomers. This type of hydrophobic knob and pocket binding seems to be essential for the interactions between BMP ligands and the type I receptors.
The latter mode of ligandreceptor interaction is characteristic for TGFβ and Activin. Both TGFβ and Activin exhibit a high affinity for type II receptors and they do not interact with the isolated type I receptors (Shi and Massague, 2003). These ligands
Trang 20bind directly to the ectodomain of the type II receptor first. This binding leads to subsequent interaction of ligandtype II receptor complex with type I receptor. This results in the formation of a huge ligandreceptor complex comprising a ligand dimer and four receptor molecules. Recent crystal structure of the human TGFβ type II receptor ectodomain and TGFβ3 complex shows that interaction occurs at the “fingertips” of the ligand dimer (Hart et al., 2002). Each monomer of the dimeric TGFβ3 interacts with one receptor. Two symmetrically positioned concave surface areas are created by this interaction, and these surface areas have been postulated to be the binding site for the ectodomain of the type I receptor.
1.2.1.3 TGF b receptors
Members of TGFβ family signal through a family of transmembrane protein serine/threonine kinases called the TGFβ receptor family (Shi and Massague, 2003). The TGFβ receptor family is classified into two subfamilies, namely the TβRI (type I receptors) and the TβRII (type II receptors) (Massague, 1998; Shi and Massague, 2003). There are 12 members found in the human TGFβ receptor family, of which seven are
glycoproteins of approximately 55 kDa and 70 kDa, respectively (Massague, 1998). Both types of receptors consist of about 500 to 570 amino acids including the signal sequence (Massague, 1998). They are organized sequentially into (a) the extracellular domain, (b) the GS domain, and (c) the kinase domain.
Generally, the extracellular domain resembles the socalled threefinger toxin fold structure (Greenwald et al., 1999). Each finger is formed by a pair of antiparallel β
Trang 21strands. The extracellular domain consists of a relatively short extracellular region, a transmembrane region and a cytoplasmic juxtamembrane region (Massague, 1998). The extracellular region comprises of 150 amino acids and is Nglycosylated. The general fold of the extracellular region is determined by the presence of 10 or more cysteines in this region. On the other hand, the transmembrane region and the cytoplasmic juxtamembrane region display no singular structural features except for (a) the Ser213 of TβRII is phosphorylated by the receptor kinase in a ligandindependent manner and this phosphorylation is required for its signalling activity, and (b) the Ser165 in the juxtamembrane region of TβRI is phosphorylated by TβRII in a liganddependent manner. The phosphorylation state of Ser165 appears to selectively modulate the intensity of different TGFβ responses.
The GS domain is an unique feature of the TβRI, but not TβRII (Massague, 1998). This domain is a highly conserved region consists of 30 amino acids and is found
The kinase domain found in both TβRI and TβRII is a typical serine/threonine protein kinase (Massague, 1998). The TβRII kinase domain is capable of
Trang 22phosphorylating TβRI on the serine and threonine residues and thus results in the activation of TβRI. In addition, the TβRII kinase domain can also phosphorylate itself. The kinase domain of TβRI phosphorylates its substrates on the serine residues only and this domain also contains a region known as L45 loop which has been known to interact with the L3 loop of receptorregulated Smad during signal transduction (Shi and Massague, 2003).
1.2.1.4 Smad proteins
The Drosophila gene product Mad (mothers against decapentaplegic) was
identified as the first intracellular mediator of TGFβ signalling (Sekelsky et al., 1995).
Subsequently, many orthologues in C.elegans and vertebrates were identified and named
as SMADs (unification of “SMA” in C. elegans and “MAD” in Drosophila) (Derynck et
al., 1996; Sekelsky et al., 1995). There are eight distinct Smad proteins identified in human, mouse, and frog (Massague, 1998; Shi and Massague, 2003). Based on their structural and functional features, SMADs are generally classified into three classes, (a) RSmads (receptorregulated Smads), (b) CoSmads (comediator Smads), and (c) I Smads (inhibitory Smads).
Among these three classes, the RSmads are the ones that are directly phosphorylated and activated by TβRI kinase (Massague, 1998; Shi and Massague, 2003). Smad 1, 2, 3, 5 and 8 form the RSmads class. Smad 1, 5 and 8 are known to respond to signalling by the BMP subfamily, and Smad 2 and 3 are the mediators of the TGFb/Activin subfamily. To date, the only known member of the CoSmads class is Smad 4 (Massague, 1998). Smad 4 associates with phosphorylated RSmads to form
Trang 23active Smad complexes. These active complexes, together with bound transporter (importin b), translocate into the nucleus to activate specific target genes transcription.
On contrary, the ISmads are a structurally divergent class of Smads (Massague, 1998; Shi and Massague, 2003) They are known to inhibit the TGFb signal transduction by competing with RSmads in binding to the phosphorylated TGFb receptors. The ISmads interaction and subsequent binding of both E3 ubiquitin ligases and the Smurfs (Smad ubiquitination regulatory factors) to the complex lead to the ubiquitination and degradation of the receptors. In human, Smad 6 and 7 form the ISmad class. The Smad
7 can inhibit both the TGFb/Activin and BMP signalling whereas Smad 6 preferentially inhibits the BMP signalling.
Generally, the Smad proteins have highly conserved Nterminal and Cterminal domains known as MH1 (MADhomology 1) domain and MH2 (MADhomology 2) domain, respectively, and an intervening linker region (Massague, 1998; Shi and Massague, 2003). The MH1 domain is made up of approximately 130 amino acids and is extremely conserved in both RSmads and CoSmad (Smad 4) but not in ISmads. In the active state, the MH1 domain displays sequencespecific DNA binding activity and may
be involved in nuclear import. In the basal state, the MH1 domain negatively regulates the function of MH2 domain by physically interacting with the MH2 domain. The MH1 domain of ISmads exhibits weak sequence homology with the MH1 domain of RSmads and it does not bind to DNA. On contrary, the MH2 domain is extensively conserved among all Smad proteins. It is about 200 amino acids long and contains receptor phosphorylation sites and a characteristic SXS motif in RSmads. It is involved in (a) receptor interaction, (b) formation of homomeric and heteromeric Smad complexes, and
Trang 24In contrast, the intervening linker region is highly diverse in length and sequence, but it contains (a) multiple phosphorylation sites which permit specific crosstalks with other existing signalling pathways, and (b) a PY motif which directs specific interaction with the Smurfs.
1.2.1.5 Recognition of Smad by the activated receptor complex
How is a specific Smad recognized by the activated receptor? From previous structural comparison and sequence analysis, it has been postulated that the basic patch adjacent to the L3 loop of RSmad is responsible for the binding of the phosphorylated
GS region of the TβRI (Wu et al., 2000). This postulation is further supported from results obtained in a study in which one of the invariant residues, His331 in the basic patch of Smad 2 was mutated and this led to a reduction in its binding affinity for phosphorylated TβRI (Huse et al., 2001). The corresponding specificity determinant in the TβRI lies in the L45 loop of the receptor kinase domain, which is found immediately next to the GS domain, and is specifically involved in the interaction with the RSmads (Chen et al., 1998; Feng and Derynck, 1997). In addition, the presence of auxillary proteins may facilitate the recognization of RSmads by the receptors. For instance, both Smad 2 and Smad 3 can be immobilized near the cell membrane surface by SARA (Smad anchor for receptor activation), through an interaction between a peptide sequence of SARA and the extended hydrophobic surface area on Smad 2 and Smad 3 (Tsukazaki et al., 1998). The SARA possesses a phospholipids binding FYVE domain which can target the molecule to the membrane of the early endosomes. The efficient recruitment of the
Trang 25Smad 2 and Smad 3 to the activated receptors for phosphorylation is facilitated by these interactions.
1.2.1.6 Mechanism of Smad phosphorylation and activation
The RSmads are directly phosphorylated by activated TβRI at 10 residues inclusive of the SXS motif present in the MH2 domain of RSmads (Shi and Massague, 2003). Phosphorylation destabilizes the interaction of Smad and SARA, and causes the Smad to dissociate from the complex. This dissociation exposes a nuclear import region
on the Smad MH2 domain. The phosphorylation of Smad also increases its affinity to CoSmad, Smad 4. The association of these two proteins, together with importin b protein, will translocate into the nucleus to elicit genes transcription.
The active form of TGFb ligand transduces its signal by first binding to type II receptor serine/threonine kinase on the cell surface. Ligand bound type II receptor serine/threonine kinase then recruits and phosphorylates the GS domain of the type I receptor serine/threonine kinase. This leads to the activation of the type I receptor serine/threonine kinase. Activated type I receptor serine/threonine kinase uses its GS domain and L45 loop to interact with the basic pocket and L3 loop of RSmad. This results in the phosphorylation of the SXS motif at the Cterminal of RSmad. The phosphorylated RSmad undergoes homotrimerization and formation of heteromeric complexes with the CoSmad. The activated Smad complexes together with bound importin b translocate into the nucleus. Inside the nucleus, the importin b dissociates
Trang 26from the complex and the translocated Smad complexes cooperate with other nuclear cofactors to form an activation complex that cooperatively binds in a precise geometry
to regulatory sequences of the target gene and initiate transcription (Shi and Massague, 2003). This TGFb signal transduction can be negatively regulated by ISmad through (i) competing with RSmads for receptor or coSmad interaction and (ii) targeting the receptors for degradation.
1.2.1.8 TGF b Subfamily and its isoforms
The TGFb subfamily has been implicated in numerous important physiological and developmental processes (Alliston et al., 2001; Cohen, 2002; Massague, 1998; Opperman
et al., 2002). There are three TGFb isoforms in mammals. They are TGFb1 (GenBank accession number BC022242), TGFb2 (GenBank accession number NM_003238), and TGFb3 (GenBank accession number NM_003239). All isoforms exhibit very similar sequence to the prototype TGFb1, particularly the active domain where the spacing of the seven cysteines is most conserved (Hinck et al., 1996). In mammals, TGFb1, TGFb2, and TGFb3 are highly conserved. Each of them has different binding affinities to their receptors and elicits different biological effects. For instance, both TGFb1 and TGFb3 are expressed in tissue structures that are undergoing morphogenesis and TGFb2 is expressed in differentiating and mature epithelium (Taipale et al., 1998).
1.2.1.9 Latent TGF b , latent TGF b binding protein, and bone
Trang 27Members of the TGFb are known to be involved in bone remodelling. They stimulate osteoid formation but inhibit mineralization process (Cohen, 2003). Latent TGFb is found abundantly in bone. The latent TGFb is a 100 kDa homodimer complex.
It consists of mature TGFb and a precursor known as LAP (latencyassociated protein). Two forms of latent TGFb, one associating with a 190 kDa LTBP (latent TGFb binding protein) and the other does not, are produced by the osteoblasts (Cohen, 2003). The associating LTBP seems to direct the latent TGFb for storage in the bone matrix. The latent TGFb can be released from the bone matrix by a matrix remodelling agent known as plasmin, which acts on the proteasesensitive hinge region of LTBP. This cleavage releases a 130 kDa portion of the LTBP and leaves the 60 kDa amino terminal portion of LTBP attached to the matrix. The tethered latent complex containing the cleaved LTBP portion and its latent TGFb attaches to the putative binding sites on the cell surface via the cleaved LTBP portion and is activated by the cellsurface associated plasmin. This activation causes the mature TGFb homodimer to dissociate from LAP to become biologically active. The biologically active TGFb homodimer binds to the TβR
II and elicits downstream signalling process. The activity and level of active TGFb is maintained by Decorin which present on the bone extracellular matrix surface and the circulating a2macroglobulin.
In contrast, latent TGFb without LTBP exists as free circulatory form of latent TGFb. The function of these latent TGFb without LTBP is unknown. It may form free pools available for activation.
Trang 281.2.1.10 Perturbation of TGF b signalling in human diseases
Perturbation of the TGFb/Smad signalling pathway underlies many human disorders such as embryonic anomalies, cancer, autoimmune disease, atherosclerosis, hypertension, osteoporosis, fibrotic diseases, hereditary hemorrhagic telangiectasia, and CamuratiEngelmann disease (Cohen, 2003; Massague, 1998). This is summarized in Table 1.
1.2.1.11 TGF b 3, and its expression patterns during early embryogenesis and in adult tissues
Three different forms of TGFb exist in the mouse and human, of which TGFb3 is
known to play a distinct role in development. Human TGF b 3 gene is located at
chromosome 14q24. It is synthesized as a prepromonomeric protein and is cleaved to form a 112 amino acid mature polypeptide (Derynck et al., 1988; ten Dijke et al., 1988a; ten Dijke et al., 1988b). TGFb3 shares approximately 80% sequence identity with TGFb1 and TGFb2 at its Cterminal 112 amino acids.
Three embryonic expression studies of Tgf b 13 were performed in mouse, of
Trang 29the lung, and some muscles (Millan et al., 1991; Pelton et al., 1991; Schmid et al., 1991a). The Tgfb3 protein is detected in cardiac muscle, heart myocytes, lens fibers, cochlear epithelium, cartilage, bone, bronchi of the lung, cuboidal epithelial lining of the kidney tubules, liver capsule, and smooth muscle of oesophagus and intestine. Table 2 is
early mouse embryogenesis.
Examination of Tgf b 3 mRNA transcripts in adult mouse tissues using Northern
Blot analysis show that Tgf b 3 transcripts are predominantly present in placenta, lung,
brain, heart, testis, adipose tissue, and male submaxillary gland (Miller et al., 1989).
Trang 30TGFb signalling components Disease TGFb TGFb
precursor a Endoglin
b
ALK1 c Type II receptor Type I
receptor SMAD 2 SMAD 3 SMAD 4 Cancer d Enhanced tumor
invasion and
metastasis
Colorectal cancer (30%) Gastric cancer (15%) Other neoplasms
Breast cancer (16%) Other neoplasms
Colorectal cancer (11%) Lung cancer (7%) Haptocellular cancer
Colorectal cancer Multiple endocrine neoplasia type
1
Pancreatic cancer (50%) Colorectal cancer (30%) Lung cancer (10%) Other neoplasms Causal
factors
Increased TGFb
expression
Somatic mutations
Somatic mutations
Somatic mutations
Loss of heterozygosity Inactivation of menin disrupts Smad 3 binding to DNA, hence block TGFb signalling
Somatic mutations
Hereditary hemorrhagic telangiectasia type 1
Hereditary hemorrhagic telangiectasia type 2
juvenile polyposis (subset) Causal
factors
TGFb
polymorphisms
Germline mutations
Germline mutations
Germline mutations
Trang 31Table 2. Expression patterns of Tgfb3 mRNA transcripts and its protein during early
+ +
Trang 321.2.1.12 TGFb3 knockout mice studies
Mouse knockout studies suggest that Tgfb3 is important in pulmonary development and palatogenesis (Kaartinen et al., 1997; Kaartinen et al., 1995;
Proetzel et al., 1995). It is found that all Tgfb3 homozygous (Tgfb3/) mutant mice
die shortly after birth and they display only two major phenotypes, abnormal pulmonary histopathology and cleft palate.
1.2.1.12.1 Abnormal pulmonary histopathology
Immediately after birth, all Tgfb3/ null pup mice present abnormal
pulmonary histopathology (Kaartinen et al., 1995). Their terminal air spaces are grossly abnormal. Their lungs have atelectic, pseudoglandular histology. The alveoli are hypoplastic, lacked of alveolar septal formation, and show mesenchymal thickening and hypercellularity. These features suggest that there is a developmental
delay in the lungs of Tgfb3/ null mice. This lung phenotype also resembles the lung
appearance of extremely premature human infants, who are at risk for developing bronchopulmonary dysplasia. In addition, extensive intrapulmonary and pleural haemorrhages are observed in these null mice.
Proetzel et al., 1995). During palatogenesis, Tgfb3 is localized in the epithelial
component of the vertical palatal shelves, predominantly in the MEE (medial edge
Trang 33epithelium) (Fitzpatrick et al., 1990). Its expression is lost as the epithelial seam
disrupts, soon after the palatal shelves fuse. In Tgfb3/ null mice, basement
membrane degradation and MEE transdifferentiation are found to be disrupted, leading to cleft palate (Kaartinen et al., 1997; Kaartinen et al., 1995; Proetzel et al., 1995).
Subsequent culture analyses performed using mouse and chick palates demonstrate that CSPG (chondroitin sulphate proteoglycan) is present on the apical surface of MEE cells immediately prior to palatal shelves adhesion (Gato et al., 2002).
This suggests that CSPG is required for palatal adhesion. It is found that Tgfb3/ null
palates lack CSPG on their MEE surface and addition of TGFb3 is able to restore palatal shelves adhesion via the stimulation of CSPG production by the MEE cells. This provides a direct role of TGFb3 for palatal shelves adhesion. It has been found that Tgfb3 can also induce filopodialike structures formation on the outer cell
membrane of wildtype (Tgfb3+/+) mouse palatal MEE prior to adhesion (Taya et al.,
1999). These filopodialike structures are absent in Tgfb3/ mouse palatal shelves.
When exogenous recombinant TGFb3 is added into culture medium of Tgfb3/
mouse palatal shelves, it can restore the fusion of these palatal shelves. In addition, Tgfb3 is known to be involved in tissue remodelling during palatal fusion (Blavier et al., 2001). Tissue remodelling during palatal fusion involves a combination of basement membrane degradation and epithelialmesenchymal transformation. Studies have shown that this remodelling process is under the control of MMPs (matrix metalloproteinases) and TIMPs (tissue inhibitors of metalloproteinases) (Blavier et
al., 2001). In the Tgfb3 null mice, the expression level of Timp2 and Mmp13 are
markedly reduced in the degrading MEE seam during fusion process (Blavier et al.,
Trang 342001). This suggests that the expression of MMPs and TIMPs is dependent on
TGFb3.
Not much is known about the molecular mechanism control of TGFb3 during
palatogenesis. However, two recent studies show that Tgfb3induced palatal fusion is mediated by Alk5/Smad2 signaling pathway (Dudas et al., 2004) and Tgfb3 activates
transcription of Lef1 (lympoidenhancing factor 1) gene in MEE cells to induce
epithelial mesenchymal transformation during mouse palate development (Nawshad and Hay, 2003).
1.2.1.13 TGFb3 mutations and diseases
1.2.1.13.1 Potential candidate gene for nonsyndromic oral clefts
TGFb3 has been proposed as one of the potential candidate genes for non
syndromic oral clefts. The strong supporting evidences are derived from previously
reported expression studies in mouse and animal knockout model. Tgfb3 is expressed
in the MEE of the palatal shelves during palatal fusion, and is required for the adhesion of the MEE and the disruption of midline epithelial seam of the palatal shelves (Fitzpatrick et al., 1990; Kaartinen et al., 1997; Kaartinen et al., 1995;
Proetzel et al., 1995). The Tgfb3 knockout mice also show developmental defect of
the secondary palate and delayed pulmonary histopathology (Kaartinen et al., 1995; Proetzel et al., 1995).
In human, the association of TGFb3 and oral clefts is controversial. Some of
the previously described population studies reported significant association of TGFb3
with nonsyndromic oral clefts in the Caucasian (Beaty et al., 2002; Lidral et al., 1998; Vieira et al., 2003), Japanese (Ichikawa et al., 2006; Sato et al., 2001) and Korean populations (Kim et al., 2003). In contrast, there are some reports that
Trang 35et al., 2004), and Japanese populations (Tanabe et al., 2000). When TGFb3 gene is
screened for the presence of mutations in the entire coding and promoter regions, no mutations are identified in the coding regions (Bayat et al., 2003). However, 21 polymorphisms are detected in the promoter region. These polymorphisms may be
important in contributing the disease pathogenesis associated with the TGFb3 gene.
1.2.1.13.2 The disease gene for ARVD1
Recently, TGFb3 has been identified as the disease gene involved in ARVD1
(arrhythmogenic right ventricular dysplasia type 1) (Beffagna et al., 2005). Three key
evidences supported the association of TGFb3 with ARVD1. First, the mapping of
ARVD1 to chromosome 14q23q24 coincides with the chromosomal location of
TGFb3 gene (Rampazzo et al., 1994). Second, two regulatory mutations (36G>A in
the 5’ UTR and 1723C>T in the 3’ UTR) in TGFb3 gene have been identified in nine
family members with proven ARVD and 40 asymptomatic members, and an unrelated proband, respectively (Beffagna et al., 2005). Third, ARVD is characterized by fibro
fatty replacement of the right ventricular myocardium, and coincidently TGFb3 is
also expressed in the myocardium
Trang 361.2.2 Zebrafish: An Animal Model for Craniofacial Development and Disease
1.2.2.1 Overview of the zebrafish model system
The zebrafish, Danio rerio, was introduced 20 years ago by George
Streisinger as a model for genetics of vertebrates (Streisinger et al., 1981). Zebrafish
is tropical freshwater fish that belong to the family of cyprinid (Detrich et al., 1999; Wixon, 2000). They are small, approximately 3cm in length, and are commonly found in the streams of India. The females are usually fat when laden with eggs and have few or none gold colouration on their undersides. On contrary, the males are slender and torpedoshaped. They have black longitudinal stripes running along their length and usually have gold colouration on their bellies and fins.
The zebrafish has 25 pairs of chromosomes and these chromosomes contain about 1.7 x 10 9 base pairs of DNA, which is about half the size of the human genome (Driever and Fishman, 1996; Postlethwait et al., 1994; Postlethwait et al., 1998).
1.2.2.2 Advantages and disadvantages of zebrafish as the animal model
For over 100 years, the laboratory mouse Mus musculus is the predominant model animal for studying genetics of development and disease. However, Mus
musculus has several disadvantages as an animal model, which includes its
intrauterine development, high maintenance cost, and it is more costly to generate knockout mice using ES (embryonic stem) cell gene targeting technology.
Since the zebrafish was introduced 20 years ago, it has emerged rapidly as an excellent model organism to bridge the gap between fly/worm and mouse/human for the study of vertebrate development and diseases. As an model system, it has several advantages (Detrich et al., 1999; Dodd et al., 2000; Penberthy et al., 2002). (1)
Trang 37Embryos develop externally and they are optically transparent. This allows easy observation of embryogenesis and organogenesis processes under a dissecting microscope. In addition, the external fertilisation enables the ease of external manipulation of the live embryos. (2) They are relatively small. Therefore, they can
be maintained in high densities in laboratory. (3) They have short generation time, approximately 23 months. This will expedite the generation of mutant fish line for study. (4) High fecundity as hundreds of embryos can be laid from each female per week. This allows large scale genetic screenings to be carried out. (5) The cost of maintaining a fish facility is relatively lower than for the mouse.
There are several known disadvantages of zebrafish as a model system. For examples, (1) The lack of ES cells for gene targeting. Therefore, knockout studies cannot be performed. (2) Duplication of zebrafish genome. This results in genetic redundancy that may complicate the comparison of homologous developmental pathways with other vertebrate species, in particular human. However, some authors consider the duplication of some genes as an advantage because it helps to understand the developmental roles of genes that cause early lethality in knockout mice.
1.2.2.3 Disease modelling in zebrafish
Presently, two approaches are available in modelling human diseases in zebrafish (Dodd et al., 2000). One approach involves analysis of mutant models derived from chemical mutagenic screens, and another involves the candidate gene approach. In the former approach, a number of mutants with phenotypes that resembles human disease states are observed. However, to identify the causative genes underlying the disease phenotypes presents a technical challenge. To date, there are 340 entries from the PUBMED search of the NCBI database
Trang 38(http://www.ncbi.nlm.nih.gov/) that use zebrafish to model human diseases. Some of these examples are illustrated in Table 3. The candidate gene approach can be used in parallel to complement the mutant models derived from mutagenic screens approach. Basically, this approach isolates and identifies zebrafish orthologues of human disease causing genes in order to model human disorders. This is usually by BLAST searches
or keyword searches containing the name of the disease against the zebrafish EST database. The disease phenotype is then established by knocking down the particular gene product
Trang 39Table 3. Zebrafish models of human disease*
Human disease Mutated gene Zebrafish
mutant
Zebrafish mutant defect
Betaspectrin reisling Anaemia. Defects in
cytoskeleton of red cell membrane
(Sun and Hopkins, 2001)
Usher 1B
syndrome
Myosin VIIA mariner Inherited deafness caused
by defects in sensory hair cell functions
Trang 40In recent years, the zebrafish has emerged as an alternative model organism to study craniofacial development and disease. Despite being quite different from mammals, many principles and features that govern craniofacial development in higher vertebrates are conserved in zebrafish (Yelick and Schilling, 2002). Cell lineage studies conducted in zebrafish show that cranial neural crest cells are formed
at the dorsal and lateral regions of the neural ectoderm at 12 hpf (Schilling and Kimmel, 1994). After gastrulation stage, they begin to migrate in three distinct streams to populate each of the seven pharyngeal arches. It is also found that each arch is made up of unique set of neural crestderived cartilages and bones that resemble those of other vertebrates (Schilling, 1997). The zebrafish cranium is made
up of six different types of cartilage, and two types of bone (dermal and cartilage replacement bone) (Benjamin, 1990; Cubbage and Mabee, 1996). Cartilage bones form the majority of bones of the zebrafish skull, approximately 43 out of 74, and they will undergo either endochondral or perichondal bone formation.
As it is, the early head skeleton in zebrafish possesses several attributes as a simple system for understanding the genetic control of craniofacial development (Kimmel et al., 2001). First, the development of the larval skeleton takes only a few days. Second, zebrafish contains craniofacial skeletal elements and muscle tissue types similar to their vertebrate counterparts. Third, their early skeletal elements are very small and mostly made of cartilages, with simple monolayered and spatial arrangements. Fourth, their craniofacial development process has been successfully examined at singlecell level in living embryos using lineagetracing techniques. High resolution fate maps are also produced and are extremely similar to those of the